SUMMARY

The first part of the thesis (Chapter 1) provides a general framework for the work presented in this thesis. The second part (Chapters 2 and 3) consisted of studies aimed at further unravelling the construct and neurobiology of impulsivity in healthy volunteers. Whether different measures of impulsivity represent largely unrelated aspects or a unitary construct in both rodents and humans was addressed in Chapter 2.  Using a within-subjects cross-species translational design, it was found that impulsive action, reflected by a failure to inhibit a prepotent response, and impulsive choice, reflected by an increased preference for immediate rewards over more beneficial delayed rewards, did not significantly correlate in both humans and rodents. In humans, a self-report measure of impulsivity represented an additional non-related aspect of impulsivity. In rodents, the within-subjects pharmacological effects of atomoxetine and amphetamine did not correlate between tasks, indicating the presence of distinct neural correlates underlying impulsive choice and impulsive action. These findings support the notion of a non-unitary nature of impulsivity and are important to acknowledge when considering the role of impulsivity in psychopathology characterised by maladaptive impulsivity and in the development of treatments targeted at improving impulse control.
In Chapter 3, evidence from different neuroimaging techniques measuring intrinsic properties of dorsal anterior cingulate cortex (dACC) functioning was combined to predict individual differences in impulsive decision making assessed by a delay discounting task. Proton MRS was used to measure glutamate concentrations in the dACC and resting-state functional activity and connectivity of the dACC was assessed as a proxy measure of spontaneous neuronal activation and communication. Individual differences in delay discounting were associated with both dACC glutamate concentrations and resting state functional connectivity of the dACC with a midbrain region including the ventral tegmental area and substantia nigra. Moreover, a mediation model demonstrated that glutamate concentrations in the dACC influenced impulsive decision making (at least partly) via functional connectivity of the dACC with the midbrain. This is the first study showing an important role for glutamate neurotransmission in human delay discounting and revealing a possible neural pathway by which dACC glutamate concentrations affect impulsive decision making. Given the critical role of impulsive decision making in addiction, our results indicate that modulation of dACC glutamate concentrations could be an important target for the treatment of substance dependent individuals.
In the third part of the thesis (Chapters 4 and 5), neurobiological and clinical effects of NAC were investigated. Preclinical evidence indicates an important role for NAC in restoring a disturbed glutamate homeostasis in rodents treated with cocaine (Baker et al. 2003b; Madayag et al. 2007) or heroin (Zhou and Kalivas 2008). Therefore, glutamate abnormalities and NAC effects on glutamate levels in the dACC, obtained using 1H MRS, were studied in Chapter 4. We found increased basal glutamate concentrations in the dACC in cocaine dependent patients compared to healthy controls.  NAC normalized glutamate concentrations in cocaine dependent patients, whereas it had no effect on glutamate concentrations in healthy controls. Moreover, consistent with the observed relationship between impulsivity and dACC glutamate concentrations in Chapter 3, higher self-reported (cognitive) impulsivity was found to be predictive of basal and NAC-induced changes in glutamate levels. The present study is the first to report findings of glutamate modulating effects of NAC in the human brain and provides a neurobiochemical rationale for implementing NAC as a treatment for cocaine dependence.
In Chapter 5, findings of a pilot study on the clinical effects of a 4-day treatment with NAC in heavy smokers were reported. Beneficial effects of NAC on the rewarding sensation of smoking and withdrawal symptoms were found. Together with one previous pilot study investigating clinical effects of NAC in heavy smokers conducted by Knackstedt et al. (2009), these results provide preliminary evidence for NAC as a promising pharmacological intervention for smoking cessation. The findings in Chapter 4 and Chapter 5 also suggest that glutamate neurotransmission is both associated with the cognitive control system (impulsivity; Chapter 4) and the motivational system (reward; Chapter 5) implicated in addiction. 
The fourth part of the thesis (Chapters 6-9) consisted of studies that focused on the neural effects of modafinil on impulse control. Although modafinil has been proposed to enhance cognitive functioning and could therefore constitute a promising pharmacological agent for the treatment of cognitive dysfunctions in psychiatric disorders, the results of the review presented in Chapter 6 warrant further investigation before implementing modafinil as a treatment strategy. The number of placebo-controlled trials is limited and the results are rather inconsistent. However, despite methodological issues such as small sample sizes, short administration periods and the lack of placebo control conditions, there is some preliminary evidence for beneficial effects of modafinil in children with ADHD, patients with a depressive disorder and cocaine dependent patients.
The studies presented in Chapter 7, Chapter 8 and Chapter 9, aimed to investigate the neural substrates of modafinil-induced modulation of impulsivity in alcohol dependent (AD) patients. With regard to impulsive action, assessed with a stop signal task, modafinil improved response inhibition only in AD subjects who showed poor initial performance on the task, whereas a deterioration in response inhibition was observed in better performing AD subjects at baseline (Chapter 7). These differential effects of modafinil on behavior was also reflected by modafinil’s effects on brain activation during successful response inhibition: activation of the supplementary motor area (SMA) and the ventrolateral nucleus of the thalamus, key regions involved in successful response inhibition, was increased in AD subjects who improved their performance under modafinil, whereas response inhibition in AD subjects who worsened after modafinil administration was associated with a modafinil-induced decrease in activation in these same brain areas. This was supported by a mediation analysis revealing that activity changes within the SMA and thalamus were largely responsible for improvements and deteriorations observed in poor and better performing AD patients, respectively. In contrast, this differential effect of modafinil in low and high impulsive AD subjects was not observed on a measure of impulsive choice (delay discounting task).
In Chapter 8, we found beneficial effects of modafinil on impulsive decision making in all AD patients, regardless of baseline performance. Importantly, this was accompanied by enhanced recruitment of frontoparietal brain regions, known to be involved in cognitive control, and decreased activation of the ventromedial prefrontal cortex, an area involved in the coding the subjective value of rewards and self-referential processes. Clearly, impulsivity is mediated by the interaction between the motivational system and the cognitive control system and modafinil improves impulsive decision making by targeting both systems. These results not only stress the important differences between independent aspects of impulsivity (Chapter 2), they also fit the inverted u-shape relationship between catecholamine neurotransmission and cognitive performance (Cools and D'Esposito 2011) which proposes that distinct optimum levels of catecholamine neurotransmitters exist for different aspects of cognitive control.
The modafinil effects on neural substrates of impulsivity presented in Chapter 7 and 8 were not limited to isolated brain regions. Instead, in both studies evidence was found that modafinil impacts the functional interaction between brain regions. Modafinil-induced improvement in response inhibition was accompanied by an increased functional coupling between the ventrolateral thalamus and the primary motor cortex, and enhanced connectivity between the superior frontal gyrus (SFG) and the ventral striatum was associated with reduced impulsive decision making after modafinil administration. These findings indicate that modafinil has a more general effect by affecting functional networks of implicated in both the motivational system and the cognitive control system (and their interaction).
In Chapter 9, we took this a step further by examining modafinil effects on the intrinsic organization of brain functioning in AD patients. To do this, the effects of modafinil on interacting large-scale resting state networks including the default mode network (DMN), salience network (SN) and central executive network (CEN), and their association with cognitive control (measured using a Stroop task). The results demonstrated that modafinil strengthens the negative functional coupling of the DMN with both the SN and CEN in AD. In addition, this increased anti-correlation between the DMN and SN and CEN was associated with modafinil-induced improvement in cognitive control in AD, indicating that modafinil modulates the functional organization and communication of the brain, which translates into enhanced cognitive control in AD.  

Taken together, the results of the presented studies can be summarized in the following main findings: (a) Impulsivity is a complex multidimensional construct consisting of several independent aspects with different underlying neural mechanisms, which is important to acknowledge when studying impulsivity in psychiatric disorders and in the development of treatments targeted at reducing impulsive behavior. (b) In addition to the neurotransmitters traditionally implicated in addiction such as dopamine, glutamate neurotransmission seems to play an important role in impulsivity and is disturbed in cocaine dependent patients. Therefore, glutamate is an important target for the treatment of addiction characterized by distorted impulse control. (c) NAC seems to be an effective and elegant agent to restore glutamate abnormalities in addiction, and these beneficial effects of NAC on glutamate abnormalities could underlie previous and current observations of beneficial clinical effects of NAC in substance dependent individuals. (d) Modafinil can enhance impulse control in alcohol dependent patients, however, the effects of modafinil on neural substrates of various measures of impulsivity are rather complex which should be acknowledged when implementing modafinil as a treatment. These main findings will be further discussed in the next section.